Astron. Astrophys. 319, 535-546 (1997)

5. H and Ca II line-profile variations

Fig. 9a shows three Ca II H-line profiles of IN Virginis. Two of them were taken in 1995 at JD 2,449,774 ( =0.922) and 2,449,781 ( =0.773), separated by approximately one rotation period, and demonstrate that there can be significant variations on a short time scale. The third observation in Fig. 9a is from 1996 at JD 2,450,103 ( =0.515), separated by approximately one year from the other two, and yet shows even stronger emission than before. So we conclude that there must be also long-term variations.

Fig. 9a and b. Line variability of chromospheric activity indicators. Panel a shows three observations of the Ca II H-emission line taken at the labeled times. It demonstrates that the emission strength varies on an annual timescale as well as on a timescale proportional to the rotation period of the star. Panel b plots our five H observations as a function of rotational phase. For comparison, a broadened and shifted spectrum of the inactive (single) star  Ser is plotted along with every IN Vir spectrum. Note that this spectrum has been cut off below intensity of 0.8 for better display.

Fig. 9b shows five H spectra as a function of rotational phase taken in 1995, one year after our Doppler image. Obviously, IN Vir exhibits a rather abnormal H line profile but not unexpected for a very active RS CVn binary. The profile shape consists of a blue-shifted emission and a narrow absorption component at the H rest wavelength. The absorption strength changes by a factor of three from phase 0.299 to phase 0.650 while the emission component apparently remains more or less constant. We note that the precision of our rotation period allows the H phase coherence to be not better than 0.13 phases per year but, nevertheless, the minimum H emission in 1995 occured at the same phase where we saw the light-curve maximum, i.e. the spot minimum, in 1994 ( 0.2-0.3) and the maximum H emission occured very near the light-curve minimum, i.e. the spot maximum. This might indicate that at least some of the profile flux in H is modulated with the rotational period and could be due to plages.

In spite of the large variations of the absorption component, no wavelength-dependent distortions or transients seem immediately obvious from the five profiles in Fig. 9b. To search for more subtle distortions due to discrete, chromospheric plages we first determined residual H spectra by subtracting the broadened and shifted spectrum of the K2-3IV star  Ser. Its spectral classification was suggested by Fekel (1996) from red-wavelength spectra and the color (see Sect. 3.4). Stars of this M-K classification are rare and are prone to chromospheric activity as soon as they rotate above, say, 5 km s^-1. The small of  Ser of 1-2 km s^-1, however, suggests it to be an inactive star and we therefore expect no or only insignificant chromospheric emission compared to IN Vir.

Fig. 10 shows the residual H profiles of IN Vir at the five rotational phases. It is immediately obvious that the profile asymmetry persists throughout all five rotational phases, with an average wavelength offset between line bisector and rest wavelength of -0.2 Å ( -10 km s^-1). This suggests a cause that is not modulated by stellar rotation and favors an interpretation with an (inhomogeneous) stellar wind model.

Fig. 10. Residual H line profiles after subtraction of an inactive reference star. The vertical, dotted line marks the H rest wavelength. Note that the profile asymmetry persists throughout all rotational phases and consequently suggests a cause that is not modulated by stellar rotation.

The similarity of the shape of the H line profile in IN Virginis to profiles seen in other active stars, e.g. in HU Virginis (Strassmeier 1994) or in HD 32918 (Vilhu et al. 1991) suggest a common cause. While HD 32918 is a single FK Comae-type star, HU Virginis (HD 106225) is a K0IV star in a single-lined spectroscopic binary with a synchronized rotation period of approximately 10 days, and thus in many respects very similar to IN Virginis. In the case of HU Vir the H profile likely results from a combination of a locally enhanced velocity field associated with two bright plage-like features apart. Strassmeier (1994) suggested a coronal loop connecting these two plages and a siphon-type mass flow within it. A similar scenario might be also the cause for the chromospheric line variability of IN Virginis, and the observed warm spot and the adjacent polar appendage in our Doppler image could be interpreted as the footpoints of a magnetic loop. However, our current H spectra have insufficient time and phase resolution to allow detailed phase-dependent profile modeling.

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998
helpdesk@link.springer.de